A novel carrier-cooperation scheme with an incentive to offer emergency lightpath support during disaster recovery

Abstract

To achieve fast recovery of optical transport networks after a disaster, we investigate a novel scheme enabling cooperation between carriers. Carriers can take advantage of their surviving or recovered optical resources to aid each other through an emergency lightpath support, thereby efficiently reducing the recovery burden, which can be heavy immediately after a disaster. Such lightpaths can be exclusively employed by counterpart carriers to satisfy their highest-priority traffic demands, such as safety confirmation and victim relief. In addition, we introduce an incentive to carriers to promote such cooperation. The carrier cooperation-planning problem is divided into eight tasks and distributed to the individual carriers and a third-party organization. During cooperation, confidential carrier information can be strictly protected by employing a carrier optical network abstraction mechanism. The evaluation results indicate that our proposed approach can significantly reduce the recovery burden and corresponding cost to carriers, resulting in a faster and more efficient disaster recovery.

Appendices

Appendix 1: CSPT constraints

The constraints on the number of available transponders are presented in (3)–(6). RWA constraints in the underlying optical network are presented in (7)–(9). The constraint on the wavelength utilization for each lightpath is at most one wavelength on a surviving long-haul fiber link, as presented in (10), including the constraints for using the co-route and the same wavelength for both directions. Equation (11) implies that wavelength utilization can also be possible if there is a restored fiber link. The degree-limitation constraint at each node is given in (12). In case of an emergency interconnection between ROADMs, there will be a limitation on the degree of ROADM. Constraints on the upper-layer packet routing are presented in (13)–(15). The constraint on the bandwidth consumption of aggregated packet traffic in lightpaths that are not for sale or purchased in carrier cooperation is presented in (16). The aggregated traffic constraints on the emergency lightpaths that are to be sold to a counterpart carrier and are candidates for purchase from such carrier during a period of cooperation are presented in (17) and (18), respectively. Equation (19) indicates that the lightpaths (i, j) ϵ Π should be created and sold to the counterpart carrier. The constraints on the border specifications are presented in (20) and (21)

$$\sum\limits_{j \in V} {\sum\limits_{{n \in V|(U_{i,n}^{w} = 0,\,{\text{or}}\,T_{i,n} \ne \inf )}} {P_{i,n}^{(i,j),w} } } \le F_{i}^{w} ,\quad \forall i \in V;\quad \forall w \in W$$

(3)

$$\sum\limits_{i \in V} {\sum\limits_{{m \in V|(U_{m,j}^{w} = 0,\,{\text{or}}\,T_{m,j} \ne \inf )}} {P_{m,j}^{(i,j),w} } } \le F_{j}^{w} ,\quad \forall j \in V;\quad \forall w \in W$$

(4)

$$\sum\limits_{j \in V} {\sum\limits_{w \in W} {\sum\limits_{{n \in V|(U_{i,n}^{w} = 0,\,{\text{or}}\,T_{i,n} \ne \inf )}} {P_{i,n}^{(i,j),w} } } } \le G_{i} ,\quad \forall i \in V$$

(5)

$$\sum\limits_{i \in V} {\sum\limits_{w \in W} {\sum\limits_{{m \in V|(U_{m,j}^{w} = 0,\,{\text{or}}\,T_{m,j} \ne \inf )}} {P_{m,j}^{(i,j),w} } } } \le G_{j} ,\quad \forall j \in V$$

(6)

$$\sum\limits_{{m \in V|(U_{m,k}^{w} = 0,\,{\text{or}}\,T_{m,k} \ne \inf )}} {P_{m,k}^{(i,j),w} } = \sum\limits_{{n \in V|(U_{k,n}^{w} = 0,\,{\text{or}}\,T_{k,n} \ne \inf )}} {P_{k,n}^{(i,j),w} } ,\quad \forall i,j,k \in V|(i \ne j \ne k);\quad \forall w \in W$$

(7)

$$\sum\limits_{{n \in V|(U_{i,n}^{w} = 0,\,{\text{or}}\,T_{i,n} \ne \inf )}} {P_{i,n}^{(i,j),w} } = v_{i,j}^{w} ,\quad \forall i,j \in V|(i \ne j);\quad \forall w \in W$$

(8)

$$\sum\limits_{{m \in V|(U_{m,j}^{w} = 0,\,{\text{or}}\,T_{m,j} \ne \inf )}} {P_{m,j}^{(i,j),w} } = v_{i,j}^{w} ,\quad \forall i,j \in V|(i \ne j);\quad \forall w \in W$$

(9)

$$\sum\limits_{i,j \in V} {\left[ {P_{m,n}^{(i,j),w} + P_{n,m}^{(i,j),w} } \right]} \le 1,\quad \forall w \in W;\quad \forall m,n \in V|U_{m,n}^{w} = 0$$

(10)

$$\sum\limits_{i,j \in V} {\left[ {P_{m,n}^{(i,j),w} + P_{n,m}^{(i,j),w} } \right]} \le \beta_{m,n} ,\quad \forall w \in W;\quad \forall m,n \in V|T_{m,n} \ne \inf$$

(11)

$$\sum\limits_{{n \in V|T_{m,n} \ne \inf }} {\beta_{m,n} + \sum\limits_{{n \in V|L_{m,n} = 1}} {L_{m,n} } } \le D_{m} ,\quad \forall m \in V|(\exists n \in V,T_{m,n} \ne \inf )$$

(12)

$$\mathop \sum \limits_{i \in V|i \ne k \ne s \ne d} \lambda_{i,k}^{s,d} = \mathop \sum \limits_{j \in V|j \ne k \ne s \ne d} \lambda_{k,j}^{s,d} , \quad \forall \left( {s,d} \right) \in R, \quad \forall k \in V$$

(13)

$$\sum\limits_{j \in V|(j \ne s \ne d)} {\lambda_{s,j}^{s,d} } = \alpha^{s,d} ,\quad \forall (s,d) \in R$$

(14)

$$\sum\limits_{i \in V|(i \ne s \ne d)} {\lambda_{i,d}^{s,d} } = \alpha^{s,d} ,\quad \forall (s,d) \in R$$

(15)

$$\mathop \sum \limits_{{\left( {s,d} \right) \in R}} \varGamma_{s,d} \lambda_{i,j}^{s,d} \le C\mathop \sum \limits_{w \in W} v_{i,j}^{w} , \quad \forall i,j \in V|\left( {i \ne j} \right)\quad {\text{and}}\quad \left( {i,j} \right) \notin \left( {\varPsi \cup \varPi } \right)$$

(16)

$$\mathop \sum \limits_{{\left( {s,d} \right) \in R}} \varGamma_{s,d} \lambda_{i,j}^{s,d} \le C\left[ {\left( {\mathop \sum \limits_{w \in W} v_{i,j}^{w} } \right) - O_{i,j} } \right], \quad \forall \left( {i,j} \right) \in \varPi$$

(17)

$$\mathop \sum \limits_{{\left( {s,d} \right) \in R}} \varGamma_{s,d} \lambda_{i,j}^{s,d} \le C\left[ {\left( {\mathop \sum \limits_{w \in W} v_{i,j}^{w} } \right) + \sigma_{i,j} } \right], \quad \forall \left( {i,j} \right) \in \varPsi - \varPi$$

(18)

$$\mathop \sum \limits_{w \in W} v_{i,j}^{w} \ge O_{i,j} ,\quad \forall \left( {i,j} \right) \in \varPi$$

(19)

$$\lambda_{s*,b}^{s*,d} \le u_{b} ,\quad \forall s* \in S,\quad \forall d \in (V - S)|(s*,d) \in R;\quad \forall b \in B$$

(20)

$$\lambda_{b,k}^{s*,d} \le u_{b} ,\quad \forall s* \in S,\quad \forall d \in (V - S)|(s*,d) \in R;\quad \forall b \in B;\quad \forall k \in V - S - B.$$

(21)

Appendix 2: TSMT constraints

The constraint on the maximum cost experienced during segment recovery task sharing/balancing by each carrier is presented in (22). The constraint assuring a task assignment in X_common for the carriers is presented in (23)

$$\mathop \sum \limits_{{\left( {i,j} \right) \in X_{\text{common}} }} p_{i,j}^{a} \gamma_{i,j}^{a} \le \lambda_{{\mathrm{max}}} ,\quad \forall a \in \Delta .$$

(22)

$$\mathop \sum \limits_{a \in \Delta } \gamma_{i,j}^{a} = 1, \quad \forall \left( {i,j} \right) \in X_{\text{common}} .$$

(23)